def get_coding(text):
"""
Function to get the coding of a text.
@param text text to inspect (string)
@return coding string
"""
for line in text.splitlines()[:2]:
try:
result = CODING_RE.search(to_text_string(line))
except UnicodeDecodeError:
# This could fail because to_text_string assume the text is
# utf8-like and we don't know the encoding to give it to
# to_text_string
pass
else:
if result:
codec = result.group(1)
# sometimes we find a false encoding that can result in errors
if codec in CODECS:
return codec
# Fallback using chardet
if is_binary_string(text):
detector = UniversalDetector()
for line in text.splitlines()[:2]:
detector.feed(line)
if detector.done:
break
detector.close()
return detector.result['encoding']
return None
def to_unicode_from_fs(string):
"""
Return a unicode version of string decoded using the file system encoding.
"""
if not is_string(string): # string is a QString
string = to_text_string(string.toUtf8(), 'utf-8')
else:
if is_binary_string(string):
try:
unic = string.decode(FS_ENCODING)
except (UnicodeError, TypeError):
pass
else:
return unic
return string